2822
J. Agric. Food Chem. 1999, 47, 2822−2829
Synthesis of Deuterated Volatile Lipid Degradation Products To Be
Used as Internal Standards in Isotope Dilution Assays. 2. Vinyl
Ketones
Jianming Lin, Dieter H. Welti, Francia Arce Vera, Laurent B. Fay, and Imre Blank*
Nestec Ltd., Nestlé Research Center, Vers-chez-les-Blanc, P.O. Box 44, CH-1000 Lausanne 26, Switzerland
The isotopically labeled compounds [5,6-2H2]-(Z)-1,5-octadien-3-one (d-I) and [1-2H1;2,2-2H1;1]-1-octen3-one (d-II), as well as the unlabeled reference compound (Z)-1,5-octadien-3-one (I) were prepared
by improved synthesis procedures. Labeling position, chemical purity, and isotopic distribution of
the compounds were characterized by various MS and NMR techniques. These molecules are used
as internal standards in quantification experiments based on isotope dilution assay. The newly
prepared compound d-II was synthesized in a simple two-step procedure, and formation of the main
isotopomers was studied in model systems.
Keywords: Synthesis; isotope labeling; deuteration; flavor compounds; lipid degradation products;
[1-2H1;2,2-2H1;1]-1-octen-3-one; [5,6-2H2]-(Z)-1,5-octadien-3-one; NMR; GC/MS; isotope dilution assay
(IDA)
INTRODUCTION
Lipid-derived volatile compounds such as (Z)-1,5octadien-3-one (I) and 1-octen-3-one (II) (Figure 1)
contribute to the characteristic and desired flavor of food
products but can also cause off-flavors depending on
their concentration. The metallic, geranium-like smelling odorant I was shown to participate in the reversion
flavor of soybean oil (Ullrich and Grosch, 1988a).
Compound II, smelling mushroom-like and metallic,
was identified as a key odorant of mushrooms (Fischer
and Grosch, 1987). These flavor compounds are formed
from linolenate (Ullrich and Grosch, 1988b) and linoleate (Grosch, 1987), respectively.
Because odorants such as I and II often occur in low
concentrations in foods, special efforts must be made
to obtain reliable quantitative results. Isotope dilution
assay (IDA) has been shown to be a powerful quantification method in flavor research [Schieberle and Grosch,
1987; review by Grosch (1994) and Schieberle (1995)].
IDA is based on the use of labeled internal standards
added to the product prior to sample preparation to
compensate for losses during the cleanup procedure. The
labeled internal standard and the analyte are monitored
by mass spectrometry (MS). Both labeled and unlabeled
flavor compounds of known purity must be available for
establishing the calibration curves.
As recently discussed by Milo and Blank (1998), the
availability of labeled compounds is still the limiting
factor in the development of IDA. The aim of this work
was to prepare deuterated analogues of I and II and
characterize their chemical and isotopic purity. The
following compounds were synthesized (Figure 1): [5,62H ]-(Z)-1,5-octadien-3-one (d-I) and [1-2H ,2-2H ]-12
1;2
1;1
octen-3-one (d-II), as well as the commercially unavailable (Z)-1,5-octadien-3-one (I).
* Author to whom correspondence should be addressed
(telephone +41/21-785-8607; fax +41/21-785-8554; e-mail
[email protected]).
Figure 1. Chemical structures of (Z)-1,5-octadien-3-one (I),
1-octen-3-one (II), and their corresponding deuterated analogues (b indicates the labeling position; the abbreviation “d”
means only that the compound is deuterated, without specifying number or position of the deuterium atoms).
EXPERIMENTAL PROCEDURES
Materials and Reagents. The unlabeled flavor compound
1-octen-3-one (II) was commercially available from Oxford
Chemicals (purity ) 98%, Brackley, England). The starting
materials for the syntheses were of highest purity: (Z)-3hexen-1-ol (IV, 95%) was from Fluka (Buchs, Switzerland),
1-octyn-3-ol (VII, 98%) was from Lancaster (Morecambe,
England), and 3-hexyn-1-ol (III, 98%) was from Aldrich (Buchs,
Switzerland). The 2H2-labeled compounds deuterium oxide
(2H2O, >99.8% 2H), lithium aluminum deuteride (LiAl2H4,
>99% 2H), and deuteromethanol (MeO2H, >99.5% 2H) were
obtained from Fluka. Deuterium gas (2H2, >99.8% 2H) was
from Carbagas (Lausanne, Switzerland) and deuterochloroform (C2HCl3, 99.8%) from Dr. Glaser AG (Basel, Switzerland).
The following reagents were used: chromium(VI) oxide (CrO3,
>99%) and pyridinium chlorochromate (PCC, g98%) from
Fluka; vinylmagnesium bromide (BrMgC2H3, 1 M in THF) and
palladium on CaCO3 (deactivated with lead, Lindlar’s catalyst)
from Aldrich.
Anhydrous dichloromethane (CH2Cl2), diethyl ether (Et2O),
pyridine, and tetrahydrofuran (THF) stored over molecular
sieves (H2O <0.005%) as well as ethanol (EtOH, >99.8%),
methanol (MeOH, >99.5%), sulfuric acid (H2SO4, 98%), and
Florisil adsorbent for chromatography (100-200 mesh) were
purchased from Fluka. Sodium hydrogencarbonate (NaHCO3),
sodium sulfate (Na2SO4), sodium acetate (NaOAc), and silica
gel 60 were from Merck (Darmstadt, Germany).
Analytical Methods. Gas Chromatography (GC). This was
performed on a Hewlett-Packard HP-5890 gas chromatograph
10.1021/jf9902090 CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/22/1999
Synthesis of Isotopically Labeled Flavor Compounds
(Geneva, Switzerland) equipped with a splitless injector and
a flame ionization detector (FID). J&W fused silica capillary
columns were employed: DB-5, DB-1701, DB-FFAP, and DBWax (30 m × 0.32 mm, film thickness ) 0.25 µm) using
different oven temperature programs (Lin et al., 1999). The
carrier gas was helium (100 kPa). Linear retention indices (RI)
were calculated according to the method of van den Dool and
Kratz (1963).
GC/MS. This was performed on a Finnigan MAT 8430 mass
spectrometer (Bremen, Germany). Electron impact (EI) mass
spectra were generated at 70 eV and positive chemical ionization (PCI) at 150 eV with ammonia as the reagent gas. Volatile
components were sampled via a Hewlett-Packard HP-5890 gas
chromatograph (Geneva, Switzerland) equipped with a cold
on-column injector. DB-1701 or DB-Wax fused silica capillary
columns were used (Lin et al., 1999). The carrier gas was
helium (90 kPa). Relative abundances of the ions are given in
percent. Selected ion monitoring (SIM) of [M + NH3]+ ions was
used for determining the distribution of doubly and triply
deuterated 1-octen-3-ol.
For the determination of isotopic purity, GC/MS was
performed on a Finnigan SSQ 7000 using PCI at 200 eV and
isobutane as reagent gas (Lin et al., 1999). The samples were
introduced by splitless injection (1 µL) using the DB-1701
capillary column described above. Helium was used as carrier
gas (70 kPa).
Nuclear Magnetic Resonance (NMR) Spectroscopy. Samples
were analyzed in WILMAD 528-PP 5 mm Pyrex NMR tubes
using deuterated chloroform as solvent. If not otherwise
specified, the NMR spectra were acquired on a Bruker AM360 spectrometer at 360.13 MHz for 1H and at 90.56 MHz for
13C under standard conditions (Lin et al., 1999). All shifts are
cited in parts per million from the internal tetramethylsilane
(TMS) standard. Proton NOE difference spectra (d-I), selective
homonuclear decoupled spectra (IV and VI), DEPT spectra (d-I
and d-II), and two-dimensional COSY (VI) and HETCOR (d-I
and VI) were acquired as described earlier (Lin et al., 1999)
to make unequivocal assignments (detailed experimental
conditions can be obtained upon request). The SPECINFO online program SPECAL (Lin et al., 1999) was applied to some
of the 13C NMR spectra for the assignment of equivocal signals.
Chemical and Isotopic Purity. This was determined on the
basis of GC-FID, GC/MS (PCI), and NMR data (Rohwedder,
1985; Lin et al., 1999). The nondeuterated substances were
analyzed by GC/MS (PCI) for isotope correction of the labeled
compounds.
Synthesis of [5,6-2H2]-(Z)-1,5-Octadien-3-one (d-I). [3,42H ]-(Z)-3-Hexen-1-ol (d-IV). A solution of 3-hexyn-1-ol (III, 4.0
2
g, 41 mmol) in MeO2H (100 mL) was deuterated with Lindlar’s
catalyst (200 mg) under normal pressure at room temperature.
Samples were taken and analyzed by GC periodically to
monitor the process of deuteration that was complete after 3.5
h. The reaction was stopped, and the catalyst was removed
by filtration. After removal of MeOH, the residue was distilled
under vacuum (51-52 °C, 12 mbar), giving rise to 3.7 g (36
mmol, 88% yield) of a colorless oil of d-IV with a purity of 98%
(GC): 1H NMR (400 MHz, C2HCl3) δ 3.62 (t, J ) 6.8 Hz, 2H,
1-CH2), 2.30 (t, J ) 6.8 Hz, 2H, 2-CH2), 2.07 (q, J ) 7.7 Hz,
2H, 5-CH2), 0.98 (t, J ) 7.7 Hz, 3H, 6-CH3); 13C NMR (100
MHz, C2HCl3) δ 134.8 (s 1:1:1, JC2H ) 23 Hz, 3-C2H), 124.4 (s
1:1:1, JC2H ) 24 Hz, 4-C2H), 62.0 (t, 1-CH2), 30.4 (t, 2-CH2),
20.4 (t, 5-CH2), 14.0 (q, 6-CH3). These spectra, compared to
those acquired with a 360 MHz instrument of commercial
nondeuterated IV (data not shown), unambiguously proved the
above molecular structure and deuteration pattern.
[3,4-2H2]-(Z)-3-Hexenal (d-V). Collin’s reagent was prepared
by adding CrO3 (25.2 g, 0.25 mol) to a stirred solution composed
of anhydrous pyridine (40 g, 0.5 mol) and CH2Cl2 (600 mL).
The mixture was stirred for 15 min at a temperature not
exceeding 25 °C. A solution of d-IV (4.2 g, 41 mmol) in
anhydrous CH2Cl2 (10 mL) was added in one portion. After
an additional 15 min of stirring at room temperature, the
solution was decanted from the residue and the latter was
washed with dry Et2O (2 × 100 mL). The combined organic
layers were filtered through a short pad of Florisil, and the
J. Agric. Food Chem., Vol. 47, No. 7, 1999 2823
filtrate was washed with cold HCl (1 N) until pH 2-3 and then
with saturated NaCl until neutral. The solvent was removed
by distillation on a Vigreux column (50 × 1 cm). The residue
was further distilled under vacuum (30-31 °C, 21 mbar),
obtaining 1.84 g (18.4 mmol, 45% yield) of d-V with a purity
of 97% (GC): GC RI(DB-5) ) 798, RI(DB-1701) ) 885, RI(FFAP) ) 1129, RI(DB-Wax) ) 1138; MS-EI 100 (34, M+), 99
(7), 86 (2), 85 (30), 84 (5), 83 (4), 82 (19), 81 (13), 80 (4), 72
(11), 71 (75), 70 (21), 69 (8), 68 (5), 58 (8), 57 (45), 56 (18), 55
(11), 54 (8), 53 (5), 52 (4), 44 (9), 43 (78), 42 (100), 41 (31), 40
(25), 39 (12); MS-CI (NH3) 118 (100, [M + NH4]+), 100 (19,
M+).
[5,6-2H2]-(Z)-1,5-Octadien-3-ol (d-VI). d-V (1.8 g, 18 mmol)
in dry THF (30 mL) was added dropwise to a stirred solution
of BrMgC2H3 (1 M, 2 0 mL) cooled at 10 °C in an ice bath. The
mixture was stirred overnight and then hydrolyzed using a
saturated solution of NH4Cl. The sample was extracted with
Et2O (2 × 50 mL). The solvent was removed, and the residue
was distilled under vacuum (105 °C, 13-14 mbar, Büchi GKR50), resulting in 1.2 g (9.4 mmol, 51% yield) of d-VI with a
purity of 93% (GC): GC RI(DB-5) ) 972, RI(DB-1701) ) 1075,
RI(FFAP) ) 1483, RI(DB-Wax) ) 1486; MS-EI 110 (2), 72 (65),
71 (23), 70 (8), 58 (8), 57 (100), 56 (7), 55 (6), 43 (12), 42 (11),
41 (5); MS-CI (NH3) 164 (1, [M + NH3 + NH4]+), 146 (7, [M +
NH4]+), 128 (100, M+ or [M + NH4 - H2O]+).
[5,6-2H2]-(Z)-1,5-Octadien-3-one (d-I). Octadienol d-VI (0.9
g, 7.0 mmol) was oxidized with PCC to the corresponding
ketone using the procedure described for the oxidation of [5,62H ]hexan-1-ol (Lin et al., 1999). About 270 mg (2.1 mmol, 30%
2
yield) of d-I was obtained with a purity of 99% (GC). The
isotopic purity of d-I was 98%, with 0.4% of [2H]-I and 1.6%
of [2H3]-I: GC RI(DB-5) ) 977, RI(DB-1701) ) 1080, RI(FFAP)
) 1364, RI(DB-Wax) ) 1371; MS-CI (NH3) 144 (100, [M +
NH4]+), 126 (20, M+).
Synthesis of (Z)-1,5-Octadien-3-one (I). (Z)-3-Hexenal
(V). In analogy to the synthesis of d-V, 3.9 g (38 mmol, 32%
yield) of V was prepared with a purity of 85% (GC) using
commercially available IV (12.0 g, 120 mmol) as starting
material: GC RI(DB-5) ) 798, RI(DB-1701) ) 891, RI(FFAP)
) 1141, RI(DB-Wax) ) 1138; MS-EI 98 (20, M+), 97 (7), 84
(18), 80 (20), 70 (11), 69 (48), 55 (28), 53 (18), 43 (39), 42 (100),
40 (40); MS-CI (NH3) 116 (100, [M + NH4]+), 98 (18, M+).
(Z)-1,5-Octadien-3-ol (VI). In analogy to the synthesis of
d-VI, 2.1 g (16.7 mmol, 60% yield) of VI was prepared with a
purity of 97% (GC) using V as starting material (2.9 g, 28
mmol): GC RI(DB-5) ) 977, RI(DB-1701) ) 1077, RI(FFAP)
) 1486, RI(DB-Wax) ) 1488; MS-EI 108 (2), 70 (70), 69 (27),
57 (100), 55 (45), 42 (17), 41 (43), 39 (15); MS-CI (NH3) 162 (1,
[M + NH3 + NH4]+), 144 (7, [M + NH4]+), 126 (100, M+ or [M
+ NH4 - H2O]+).
(Z)-1,5-Octadien-3-one (I). The octadienol VI (2.0 g, 16 mmol)
was oxidized to I in the same way as described for the oxidation
of [5,6-2H2]hexan-1-ol (Lin et al., 1999). About 370 mg (3.0
mmol, 19% yield) I was obtained with a purity of 95% (GC):
GC RI(DB-5) ) 979, RI(DB-1701) ) 1085, RI(FFAP) ) 1365,
RI(DB-Wax) ) 1371; MS-CI (NH3) 142 (100, [M + NH4]+), 124
(28, M+).
Synthesis of [1-2H1;2,2-2H1;1]-1-Octen-3-one (d-II). [12H ,2-2H ]-1-Octen-3-ol (d-VIII). In a 200 mL three-neck
1;2
1;1
reactor fitted with a reflux condenser and a thermometer was
suspended 2.0 g (48 mmol) of LiAl2H4 in anhydrous THF.
Compound VII (5.0 g, 40 mmol) dissolved in anhydrous THF
(10 mL) was slowly added to the magnetically stirred solution.
The mixture was refluxed under nitrogen. GC analysis indicated complete reduction of VII to d-VIII after 18 h. On cooling
in an ice bath, 20 mL of heavy water was added drop by drop,
followed by 80 mL of aqueous H2SO4 (4 N) to dissolve any
insoluble residues. The organic phase was separated from the
water phase, and the aqueous solution was then extracted with
Et2O (3 × 50 mL). The combined organic solutions were
washed successively with saturated solutions of NaHCO3 (2
× 10 mL) and NaCl (2 × 10 mL) and then dried over
anhydrous Na2SO4. After removal of the solvent by evaporation
and distillation under vacuum (59-61 °C, 10 mbar), 4.8 g (37
mmol, 92% yield) of a colorless oil of d-VIII with a purity of
2824 J. Agric. Food Chem., Vol. 47, No. 7, 1999
Lin et al.
Figure 2. Mass spectrum (EI) of [5,6-2H2]-(Z)-1,5-octadien-3-one (d-I, A) and (Z)-1,5-octadien-3-one (I, B) (b indicates the labeling
position).
Scheme 1. Synthesis of
[5,6-2H2]-(Z)-1,5-Octadien-3-one (d-I) and
(Z)-1,5-Octadien-3-one (I)a
a b indicates the labeling position; the abbreviation “d”
means that the compound is deuterated, without specifying
number or position of the deuterium atoms.
97% (GC) was obtained: GC RI(DB-5) ) 984, RI(DB-1701) )
1076, RI(FFAP) ) 1545, RI(DB-Wax) ) 1456; MS-EI 113 (1),
112 (1), 100 (2), 99 (10), 88 (10), 87 (9), 85 (4), 84 (7), 83 (10),
82 (5), 76 (2), 75 (19), 74 (18), 71 (7), 70 (8), 69 (8), 68 (5), 61
(7), 60 (100), 59 (84), 58 (8), 57 (12), 56 (8), 55 (12), 44 (6), 43
(19), 42 (5), 41 (11); MS-CI (NH3) 149/148 (40/48, [M + NH4]+),
131/130 (100/85, M+ or [M + NH4 - H2O]+).
[1-2H1;2,2-2H1;1]-1-Octen-3-one (d-II). Oxidation of d-VIII
(4.75 g, 36.5 mmol) was performed as described for the
oxidation of [5,6-2H2]-hexan-1-ol (Lin et al., 1999) using 11.8
g (55 mmol) of PCC and 0.9 g (11 mmol) of anhydrous NaOAc
suspended in anhydrous CH2Cl2. Distillation under vacuum
(56-60 °C, 16 mbar) gave ∼1.9 g of a yellow oil with an intense
metallic smell. The oil was further purified by column chromatography on silica gel by eluting first with pentane (100
mL) and then with pentane/Et2O (95 + 5, v/v), resulting in
1.6 g (12.2 mmol, 34% yield) of compound d-II with a purity
of 99% (GC). d-II was composed of [2H2]-II (55%), [2H3]-II (44
%), and [2H]-II (1%): GC RI(DB-5) ) 973, RI(DB-1701) ) 1064,
RI(FFAP) ) 1291, RI(DB-Wax) ) 1301; MS-CI (NH3) 147/146
(100/90, [M + NH4]+), 130/129 (4/4, [M + H]+).
Reduction of 1-Octyn-3-ol (VII) with Different Amounts
of LiAl2H4. 1-Octyn-3-ol (VII, 0.6 g, 5 mmol) and various
amounts of LiAl2H4 were refluxed in 10 mL of anhydrous THF
as described aboved in the preparation of d-VIII. The molar
ratios of LiAl2H4 to VII were 1:4, 1:2, 1:1, and 2:1. Aliquots of
the reaction mixtures were taken after refluxing for 0.5, 1, 2,
4, and 18 h. The reaction products were first hydrolyzed with
2H O (1 mL) and then treated with aqueous H SO (4 N, 2
2
2
4
mL). The aqueous layers were extracted with Et2O. The solvent
extracts were dried over Na2SO4 and then diluted for GC-FID
and GC/MS analysis. The solvent was completely evaporated
for NMR measurements.
RESULTS AND DISCUSSION
Synthesis and Characterization of [5,6-2H2]-(Z)1,5-Octadien-3-one (d-I) and (Z)-1,5-Octadien-3-one
(I). Compounds d-I and I were prepared as shown in
Scheme 1 by adapting the method reported by Ullrich
and Grosch (1988b) and Guth and Grosch (1990).
Lindlar’s catalyst was used for partial deuteration of
3-hexyn-1-ol (III) to [3,4-2H2]-(Z)-3-hexen-1-ol (d-IV) via
stereospecific syn-addition. According to Rakoff and
Rohwedder (1992), derivatization of the OH group in
III to eliminate replaceable hydrogen is not required
prior to the deuteration step. Oxidation of d-IV/IV was
carried out with Collin’s reagent to avoid isomerization
of d-V/V due to the reagent’s lower acidity compared to
PCC (Kajiwara et al., 1975). Chain elongation was
achieved by a Grignard reaction of the aldehydes (d-V/
V) with vinylmagnesium bromide followed by oxidation
of the resulting alcohols (d-VI/VI) with PCC.
The MS-EI data of the intermediate reaction product
VI were identical with those of (Z)-1,5-octadien-3-ol
reported by Tressl et al. (1982). MS-CI of d-VI indicated
incorporation of two deuterium atoms. The mass spectra
of d-I and its unlabeled analogue (I) are shown in parts
A and B of Figure 2, respectively. These data were in
good agreement with those reported by Guth (1991) and
Swoboda and Peers (1977), respectively. The shift of two
units in d-I was indicated by the fragments at m/z 111,
97, 83, and 71. MS-CI of d-I confirmed incorporation of
two deuterium atoms.
The 1H NMR data of VI, I, and d-I are summarized
in Table 1 and the 13C NMR data of VI and d-I in Table
2. Spectral interpretation of VI helped to assign the
NMR signals of I and d-I. Besides the exchangeable
hydroxyl proton at 1.7 ppm, the proton spectrum of VI
(Figure 3) represented nine different single spins or
groups of magnetically equivalent spins forming a
coupling network not easily separated into subgroups.
Gaussian resolution enhancement was applied to the
free induction decay to better resolve the complex line
patterns (resolution enhanced spectrum not shown). A
COSY spectrum permitted unambiguous assignment of
both ends of the molecule. However, selective homonuclear decoupling experiments were required for unequivocal attribution of the 5-CH and 6-CH signals.
Irradiation of the 4-CH2 signal produced a loss of
doublet coupling of ∼7.5 Hz on the 5-CH signal and a
loss of triplet coupling of ∼1.5 Hz on the 6-CH signal,
whereas irradiation of the 7-CH2 signal produced a loss
of doublet coupling of ∼7.5 Hz on the 6-CH signal and
a loss of triplet coupling of ∼1.5 Hz on the 5-CH signal.
The coupling of 10.9 Hz between 5-CH and 6-CH
demonstrates cis configuration of the double bond.
Carbon assignments were derived from those of the
protons via a one-bond HETCOR experiment.
Proton NMR data of I (Table 1) obtained at 400 MHz
are in reasonable agreement with those reported by
Swoboda and Peers (1977), who were, however, not able
to distinguish the 5-H and 6-H signals at 90 MHz
spectral frequency. No NMR data for d-I were found in
Synthesis of Isotopically Labeled Flavor Compounds
Table 1.
1H
J. Agric. Food Chem., Vol. 47, No. 7, 1999 2825
NMR Data of (Z)-1,5-Octadien-3-ol (VI), (Z)-1,5-Octadien-3-one (I), and [5,6-2H2]-(Z)-1,5-Octadien-3-one (d-I)a
proton
(Z)-1,5-octadien-3-ol (VI)
(Z)-1,5-octadien-3-one (I)b
1-CHcis
1-CHtrans
2-CH
3-H
3-OH
5.26, d t, J ) 17.2, 1.5 Hz, 1 H
5.13, d d dc, J ) 10.5, 1.5, 1.3 Hz, 1 H
5.90, d d d, J ) 17.2, 10.5, 5.8 Hz, 1 H
4.15, d t tc, J ) 6.8, 5.8, 1,4 Hz, 1 H
1.7, broad, >1 H (includes fast exchanging
H2HO protons from solvent)
2.32, complex multiplet, g25 linesc, 2 H
5.36, d t t, J ) 10.9, 7.5, 1.6 Hz, 1 H
5.58, d t t, J ) 10.9, 7.3, 1.5 Hz, 1 H
2.08, quintet of d tc, J ) 7.5, 1.6, 0.7 Hz, 1 H
0.97, t, J ) 7.5 Hz, 3 H
6.26, d d, J ) 18.0, 1.5 Hz, 1 H
5.84, d d, J ) 10.5, 1.5 Hz, 1 H
6.38, d d, J ) 18.0, 10.5 Hz, 1 H
6.26, d d, J ) 17.6, 1.4 Hz 1 H
5.84, d d, J ) 10.4, 1.4 Hz, 1 H
6.39, d d, J ) 17.6, 10.3 Hz, 1 H
3.35, d, J ) 7.0 Hz, 2 H
5.54d, m, 1 H
5.62d, m, 1 H
2.08, quintet, J ) 7.5 Hz, 2 H
0.98, t, J ) 7.6 Hz, 3 H
3.35, s (slightly broad), 2 H
-e
-e
2.07, q (slightly broad), J ) 7.5 Hz, 2 H
0.99, t, J ) 7.5 Hz, 3 H
4-CH2
5-CH
6-CH
7-CH2
8-CH3
[5,6-2H2]-(Z)-1,5-octadien-3-one (d-I)
a Chemical shift δ in ppm from internal TMS. The multiplicity abbreviations used to describe NMR signals are s (singlet), d (doublet),
t (triplet), q (quartet), and m (multiplet). For the proton spectra, d t means doublet of triplets, with decreasing values of the absolute
coupling constants (here: |J|doublet > |J|triplet). b Spectrum measured on a Bruker 400 MHz instrument in C2HCl3 with TMS as internal
standard (room temperature, not precisely specified). Shifts and coupling constants for the 400 MHz spectra were determined graphically
and may be less precise than the 360 MHz data. c Gaussian resolution enhancement required to resolve these lines. d Assignments not
verified, might have to be interchanged. e The total integral value in the range 5.7-5.4 ppm was ∼0.04 H. These trace signals are not
necessarily due to residual 5-CH and/or 6-CH of the title compound, but may also be caused by other impurities.
Table 2. 13C NMR Data of (Z)-1,5-Octadien-3-ol (VI) and
[5,6-2H2]-(Z)-1,5-Octadien-3-one (d-I)a
carbon
1-CH2
2-CH
3-CdO
3-C-OH
4-CH2
5-CH
5-C2H
6-CH
6-C2H
7-CH2
8-CH3
(Z)-1,5-octadien3-ol (VI)
114.7, t
140.5, d
72.4, d
35.0, t
123.7, d
135.3, d
20.7, t
14.2, q
[5,6-2H2]-(Z)-1,5-octadien3-one (d-I)
128.5, t
136.0, d
198.6, s
38.5, t
119.6, s, 1:1:1, 1JC2H ) 24.5 Hz
135.0, s, 1:1:1, 1JC2H ) 23.6 Hz
20.8, t
13.9, q
a For 13C NMR spectra, s, d, t, and q denominate quaternary,
CH, CH2, and CH3 carbons, respectively.
the literature. All of the signals of d-I could be individually assigned. On the basis of the spectrum of I, the
resonances of the residual 5-H and 6-H protons would
be expected in the region of 5.7-5.4 ppm. Among the
spurious small signals in this region, no resonances
Figure 3.
1H
clearly displaying the expected characteristics were
found. Because of the deuteration, the (Z)-character of
the 5/6-C double bond could not be demonstrated from
the coupling constants. NOE difference spectra (irradiation of 4-CH2) yielded a small NOE on the 7-CH2 signals,
supporting the cis arrangement. Most carbon signals of
d-I were assigned via correlation to the proton signals
by a two-dimensional HETCOR experiment. The two
deuterated carbons were assigned by shift prediction
with the SPECINFO SPECAL program and by comparison with VI.
Synthesis and Characterization of [1-2H1;2,22H ]-1-Octen-3-one (d-II). Compound d-II was pre1;1
pared in a newly developed simple two step procedure
(Scheme 2) by LiAl2H4 reduction of 1-octyn-3-ol (VII)
to deuterated 1-octen-3-ol (d-VIII) followed by subsequent oxidation with PCC following the general working
procedures reported by Bates et al. (1954), Snyder
(1967), and Corey and Suggs (1975). This synthesis
route was chosen to obtain stable molecules that are
not susceptible to 2H/H exchange provoked by enoliza-
NMR spectrum of (Z)-1,5-octadien-3-ol (VI) with multiplet expansions (for explanation, see text).
2826 J. Agric. Food Chem., Vol. 47, No. 7, 1999
Lin et al.
Scheme 2. Synthesis of a Mixture of
[1,2-2H2]-1-Octen-3-one and [1,1,2-2H3]-1-Octen-3-one
(d-II)a
a
b indicates the labeling position; the abbreviation “d”
means that the compound is deuterated, without specifying
number or position of the deuterium atoms.
Figure 4. Mass spectrum (EI) of d-II composed of a mixture
of [1,2-2H2]-1-octen-3-one and [1,1,2-2H3]-1-octen-3-one (b indicates the labeling position without specifying the number
of deuterium atoms).
tion. Such phenomena have been observed for the
compound [4,5-2H2]-1-octen-3-one (Guth and Grosch,
1990).
The presence of 1-octen-3-one was indicated by the
RI values on different stationary phases. The MS-CI
data suggested d-II to be a mixture of doubly and triply
deuterated 1-octen-3-ones with nearly equivalent distribution. The shift of two or three units was indicated
in the MS-EI spectrum of d-II (Figure 4) by the
fragments at m/z 99/100, 85/86, 72/73 (McLafferty
rearrangement), and 57/58.
1H NMR and 13C NMR data of d-II are shown in
Table 3. The assignment of the aliphatic part was
straightforward. The olefinic protons were identified by
comparison with the proton spectrum of I. Our results
for d-II agree well with earlier 1H NMR data on II
Table 3.
1H
proton
1-CHcisc
1-CHtransc
2-CH
NMR and
13C
Figure 5. Structures of various isotopomers of d-II identified
by NMR spectroscopy [the rest R represents -CO-(CH2)4CH3].
reported by Swoboda and Peers (1977) and with 13C
NMR shifts of the nonlabeled compound given by
Pritzkow et al. (1979). The integral values of the olefinic
protons demonstrated that d-II contained about half
[1,1,2-2H3]-1-octen-3-one (d-IIA in Figure 5) and half
[1,2-2H2]-1-octen-3-one; the deuterium atom at 1-C was
nearly equally distributed between the cis (d-IIB) and
trans (d-IIC) isomers. Small percentages of cis and trans
[1-2H]-1-octen-3-one were also present, that is, d-IIE and
d-IIF (Figure 5). The NMR spectra in the olefinic region
are rationalized as discussed in the following.
1H NMR (Figure 6A) revealed signals at ∼6.2 ppm
generated by 1-CH protons in the cis position on the
double bond with respect to the rest of the molecule.
The three bigger lines are equidistant with a splitting
of 2.6 Hz (d-IIB). They represent a 1:1:1 signal with
additional line broadening probably due to an unresolved small deuterium splitting (the fact that the
middle line is higher than the outer ones is caused by
their incomplete separation). Therefore, it can be postulated that this signal is due to a double bond substituted with two deuterium atoms, where 3JH2H is 2.6 Hz
and 2JH2H much smaller. The two small “satellite” lines
have a distance of 17.6 Hz, typical for a trans vinyl
3J
HH and thus are due to a single cis proton coupling to
a 2-CH (d-IIE).
The same reasoning applies for the signal at 5.8 ppm,
which represents 1-CH protons trans to the rest of the
molecule, with 3JH2H ) 1.6 Hz (d-IIC) and 3JHH ) 10.7
Hz (d-IIF). The integrals of the signal groups near 6.2
NMR Data of [1-2H1;2,2-2H1;1]-1-Octen-3-one (d-II)a (for Assignment of Isotopomers, See Text)
1H
NMRb
6.20, ∼1:1:1, J ) 2.6 Hz (d-IIB) and d (small),
J ) 17.6 Hz (d-IIE), total 0.2-0.3 H d
5.80, ∼1:1:1, J ) 1.6 Hz (d-IIC) and d (small),
J ) 10.7 Hz (d-IIF), total 0.2-0.3 H d
6.3-6.4 (several signals superimposedf,
d-IID, d-IIE, d-IIF)
carbon
1-CH2H
1-C2H2
2-CH
2-C2H
4-CH2
5-CH2
6,7-CH2
2.58, t, J ) 7.5 Hz, 2 H
1.63, m (quintet), Jav ) 7.5 Hz, 2 H
1.2-1.4, complex m, 4 H
8-CH3
0.90, t, J ) 6.9 Hz, 3 H
3-CdO
4-CH2
5-CH2
6-CH2
7-CH2
8-CH3
13C
NMR
127.49e, d (∼1:1:1), 1JC2H ) 24.4 Hz (d-IIB or d-IIC),
127.46e, d (∼1:1:1), 1JC2H ) 24.5 Hz (d-IIB or d-IIC)
127.2, s (probably 1:2:3:2:1), 1JC2H 24.4 Hz (d-IIA,
overlapping with the 1-CH2H signals)
136.56, 136.54, 136.47, all d, all low intensity (d-IID-F)
136.23g, s, 1:1:1, 1JC2H ) 24.6 Hz (d-IIB or d-IIC)
136.22g, s, 1:1:1, 1JC2H ) 24.5 Hz (d-IIB or d-IIC)
136.14g, s, 1:1:1, 1JC2H ) 24.5 Hz (d-IIA)
∼201.14, s, g3 unresolved lines
39.60, t
23.72h, t
31.46, t
22.50h, t
13.94, q
Chemical shift δ in ppm from internal TMS. The multiplicity abbreviations used to describe 1H NMR signals are s (singlet), d (doublet),
t (triplet), q (quartet), and m (multiplet). For 13C NMR spectra, s, d, t, and q denominate quaternary, CH, CH2, and CH3 carbons, respectively.
b Interpulse spacing ≈ 220 s. c Cis and trans refer to the proton position at 1-C with respect to the aliphatic rest of the molecule at 2-C.
d Superimposed signals of at least two isotopomers. e Assignment to either cis or trans proton isotopomer not attempted. f Signals correspond
to 1-CH2, 1-CH2H, and 1-C2H2 isotopomers with total <0.05 H. g For assignment of these signals, see text. h Most likely assignment for
5-CH2 and 7-CH2 based on SPECAL spectra prediction.
a
Synthesis of Isotopically Labeled Flavor Compounds
J. Agric. Food Chem., Vol. 47, No. 7, 1999 2827
Table 4. Formation of Deuterated 1-Octen-3-ol (d-VIII)
by Reduction of 1-Octyn-3-ol (VII) with Different
Amounts of LiAl2H4 at Various Reflux Timesa
ratio LiAl2H4/VII
sample
1
2
3
4
molar basis
2H/H
acidic
1:4
1:2
1:1
2:1
1:2
1:1
2:1
4:1
reflux time
0.5 h
1h
2h
4h
18 h
0
50
60
30
0
50
70
35
0
50
80
40
0
50
90
50
0
50
100
100
a Approximate yields (in percent) were obtained by GC analysis.
Abbreviation “d” only means that the compound is deuterated,
without specifying number or position of the deuterium atoms.
Figure 6. NMR spectra of the olefinic region of deuterated
1-octen-3-one (d-II) with expansions (for explanation, see
text): (A) 1H NMR; (B) 13C NMR; (C) 13C-DEPT.
and 5.8 ppm amount to ∼0.25 proton values each. The
complex residual signal between 6.4 and 6.3 ppm
amounts to ∼0.05 proton equivalents and stands for
byproduct isotopomers (d-IID, d-IIE, and d-IIF) bearing
a proton instead of a deuterium atom in the 2-position.
The olefinic part of the 13C NMR spectrum is shown
in Figure 6B. The dominant 1-C signals are two sets of
doublet 1:1:1 line patterns near 127.5 ppm, which
correspond to [1,2-2H2]-II with deuterium at either the
cis (d-IIB) or the trans (d-IIC) position of 1-C. For [1,1,22H ]-II (d-II ), a singlet 1:2:3:2:1 line pattern near 127.2
3
A
ppm with lower intensity would be expected (upfield
one-bond isotope effect of ∼0.3 ppm). Because of overlap
with the 1:1:1 patterns, only its two rightmost lines can
be observed. Note that the 13C relaxation behavior of
different isotopomers can be very different, so that these
signals should not be used for quantification. With
sufficient digital resolution (i.e., by zero-filling of the
free induction decay), three overlapping dominating
1:1:1 patterns can be found in the 2-C region near 136.2
ppm. They represent d-IIA, d-IIB, and d-IIC, which do
not show up in the DEPT spectrum (Figure 6C). As for
1-C, we expect an upfield two-bond deuterium isotope
effect of ∼0.1 ppm. Therefore, the 2-C 1:1:1 signal at
highest field is tentatively assigned to d-IIA, whereas
the hardly resolved two signal groups at lower field can
be attributed to d-IIB and d-IIC. As for 1-C, it is not
easy to say which signal belongs to cis and which to
trans.
The three downfield doublet signals appearing in the
DEPT spectrum (Figure 6C) are due to three more
isotopomers containing a 2-CH proton instead of a
2-C2H deuterium (d-IID, d-IIE, and d-IIF). Because they
occur in very small quantities only, they are not further
discussed. The absence of negative (triplet) signals near
127.5 ppm in the DEPT spectrum indicates that isotopomers bearing two terminal vinyl protons do not occur
in appreciable amounts.
Formation Mechanism of [1-2H1;2,2-2H1;1]-1-Octen3-one (d-II). In the synthesis of d-II, deuterium was
introduced by reduction of 1-octyn-3-ol (VII) with LiAl2H4
(Scheme 2). A set of experiments was carried out to
study the formation of doubly and triply deuterated
1-octen-3-ol, that is, [1,2-2H2]-VIII and [1,1,2-2H3]-VIII.
Two aspects were of special interest: the overall yields
of labeled 1-octen-3-ol obtained and the distribution of
isotopomers formed.
Reduction of VII to deuterated 1-octen-3-ol (d-VIII)
was monitored by GC analysis (Table 4). In sample 1,
the reaction did not result in any d-VIII when using
LiAl2H4/VII in a ratio of 1:4. Partial reduction of VII
was observed by increasing the ratio LiAl2H4/VII to 1:2,
and about half of the starting material was reacted to
1-octen-3-ol after 0.5 h (sample 2). However, further
reflux time did not increase yields of the alcohol,
indicating that the amount of LiAl2H4 was not sufficient.
For a complete reduction, the molar ratio of starting
material to reagent should be at least 1:1 (sample 3).
1-Octen-3-ol was then continuously generated over time
and the reaction was completed after 18 h. The reaction
was found to be slower when using an excess of LiAl2H4;
that is, yields of d-VIII were only 40% after 2 h, whereas
80% was produced in the presence of equimolar amounts
of LiAl2H4 and alkynol (sample 4).
GC/MS analysis of sample 1 revealed that the starting
material 1-octyn-3-ol (VII) was transferred into the
monodeuterated analogue (d-VII). The exact labeling
position was determined by 1H NMR (δ/ppm): VII
showed characteristic signals at 2.47 (d, J ) 2.2 Hz, 1
H, 1-CH) and 4.37 (broad t d, J ) 6.6, 1.8 Hz, 1 H,
3-CH), and d-VII was characterized by signals at 2.47
(d, J ) 2.2 Hz, 0.04 H, 1-CH) and 4.37 (broad t, J ) 6.4
Hz, 1 H, 3-CH). In both VII and d-VII, the 3-CH signals
were broad, so that coupling constant could not exactly
be determined. To obtain quantitative proton signals,
relaxation delays of at least 80 s had been used. The
2828 J. Agric. Food Chem., Vol. 47, No. 7, 1999
Lin et al.
Scheme 3. Schematic Formation Mechanism of
[1-2H]-1-Octyn-3-ol, [1,2-2H2]-1-Octen-3-ol, and
[1,1,2-2H3]-1-Octen-3-ol by Reacting LiAl2H4 with
1-Octyn-3-ol (R ) C5H11) at Different Molar Ratiosa
Figure 7. Formation of [2H3]-1-octen-3-ol ([2H3]-VIII) and
[2H2]-1-octen-3-ol ([2H2]-VIII) by reacting LiAl2H4 with 1-octyn3-ol (VII) at the LiAl2H4/VII molar ratios of 1:2, 1:1, and 2:1
for various reflux times monitored by GC/MS.
signal intensity of the acetylenic proton in d-VII was
reduced to 0.04 H and the signal multiplicity of the
proton attached to C-3 changed from a triplet of doublets
to a triplet pattern. On the basis of these data, it was
concluded that the deuterium atom was located at the
terminal alkyne carbon (1-C), resulting in ∼96% of
[1-2H]-VII. Alkynic monodeuteration was also confirmed
by 13C NMR spectra (data not shown).
The relative distribution of doubly or triply deuterated
1-octen-3-ol was analyzed by GC/MS in the CI mode
using NH3 as reagent gas by monitoring the [M + NH4]+
ions. Ratios of [2H3]-VIII to [2H2]-VIII were obtained
after isotope correction (Rakoff and Rohwedder, 1992).
As shown in Figure 7, doubly deuterated 1-octen-3-ol
was exclusively obtained when using a LiAl2H4/VII
molar ratio of 1:2. When the LiAl2H4/VII molar ratio
was raised to 1:1 or 2:1, the amounts of triply deuterated
1-octen-3-ol increased with reaction time up to 40 and
50% of the total composition, respectively. A slight
excess of LiAl2H4 over VII (molar ratio of 1.2:1, conditions described under Experimental Procedures) also
resulted in d-II containing about equal proportions of
doubly and triply deuterated isotopomers.
The formation of different isotopomers of II may be
explained by the structural particularity of the R-alkynol
used as the starting material. In general, metal hydrides
preferably react with “active” hydrogen atoms, liberating hydrogen gas. In VII there are two distinct regions
with active hydrogen atoms, that is, the hydroxyl group
and the acetylenic hydrogen. Experimental findings
suggest that the favored reaction at a LiAl2H4/VII molar
ratio of 1:4 was the replacement of the acetylenic
hydrogen (sample 1 in Table 4), thus forming [1-2H]VII upon deuterolysis (pathway 1 in Scheme 3). As
shown by Mole and Surtees (1963), the acetylenic
deuterium is stable toward 2H/H exchange under the
conditions of hydrolysis. In agreement with that, additional washing of a sample containing [1-2H]-VII and
unlabeled 1-octyn-3-ol (VII) with aqueous acid did not
change the 2H/H ratio (data not shown).
At a LiAl2H4/VII molar ratio of 1:2 (sample 2), the
formation of [2H2]-1-octen-3-ol can be explained by
attack of LiAl2H4 at the hydroxyl group (pathway 2 in
Scheme 3). This reaction corresponds to reduction of
R-alkynols to olefinic alcohols (Bates et al., 1954). The
preferred formation of trans-configured olefinic alcohols
may be due to geometrical constraints of the cyclic
intermediate postulated by Snyder (1967), which rapidly
undergoes deuteride transfer to the C-2 acetylenic
carbon. Subsequent deuterolysis results in [1,2-2H2]VIII in 50% yields (Table 4), whereas the remaining
a In the reaction intermediates, aluminum may bind to
deuterium atoms (2H) and/or to the starting molecule 1-octyne3-ol via Al-O or Al-C bonds, depending on the molar ratio of
LiAl2H4 to 1-octyn-3-ol.
50% of the reaction mixture was [1-2H]-VII formed via
reaction pathway 1.
[2H3]-VIII was generated in the presence of an excess
of active 2H compared to active H (samples 3 and 4 in
Table 4) favoring the attack of LiAl2H4 at both active
hydrogens in VII. Deuteride transfer and deuterolysis
led to [1,1,2-2H3]-1-octen-3-ol as shown in pathway 3
(Scheme 3).
These results suggest that [1,1,2-2H3]-1-octen-3-ol can
be obtained with high isotopomeric purity by first
treating VII with LiAl2H4 (ratio 4:1) followed by complete reduction of [2H]-VII to [2H3]-VIII with an excess
of LiAl2H4. Alternatively, [1,2-2H2]-1-octen-3-ol can be
obtained by reacting LiAl2H4 and VII in the molar ratio
of 1:2. However, the maximal yield is 50% and removal
of residual VII is required.
Conclusions. Two labeled vinyl ketones and one
nonlabeled vinyl ketone were prepared by applying new
or modified synthesis procedures. Their yields as well
as chemical and isotopic purity were characterized using
MS and NMR techniques. Mixtures of several isotopomers were obtained in the synthesis of deuterated
1-octen-3-one. However, this does not preclude its use
as an internal standard in IDA. In view of the likely
future demand for such labeled reference compounds,
more work on optimization of synthesis procedures from
a yield, time, and cost perspective may be justified.
Synthesis of Isotopically Labeled Flavor Compounds
ACKNOWLEDGMENT
We acknowledge Mr. R. Fumeaux for helpful discussions and expert technical assistance in some of the
syntheses, as well as Mrs. S. Metairon for her expertise
in GC/MS analysis. We thank Dr. E. Prior for linguistic
proofreading of the manuscript and Mrs. J. Lindstrand
for database literature searches.
LITERATURE CITED
Bates, E. B.; Jones, E. R. H.; Whiting, M. C. Research on
acetylenic compounds. Part XLII. Reduction with lithium
aluminium hydride. J. Chem. Soc. 1954, 1854-1860.
Corey, E. J.; Suggs, J. W. Pyridinium chlorochromate. An
efficient reagent for oxidation of primary and secondary
alcohols to carbonyl compounds. Tetrahedron Lett. 1975, 31,
2647-2650.
Fischer, K.-H.; Grosch, W. Volatile compounds of importance
in the aroma of mushrooms (Psalliota bispora). Lebensm.
Wiss. Technol. 1987, 20, 233-236.
Grosch, W. Reactions of hydroperoxidessProducts of low
molecular weight. In Autoxidation of Unsaturated Lipids;
Chan, H. W.-S., Ed.; Academic Press: London, U.K., 1987;
pp 95-139.
Grosch, W. Determination of potent odourants in foods by
aroma extract dilution analysis (AEDA) and calculation of
odour activity values (OAVs). Flavour Fragrance J. 1994,
9, 147-158.
Guth, H. Verderb von Sojaöl unter Einwirkung von Licht und
Sauerstoff - Identifizierung der Aromastoffe und Vorläufer.
Ph.D. Dissertation, Technical University of Munich, Germany, 1991 (in German).
Guth, H.; Grosch, W. Deterioration of soya-bean oil: Quantification of primary flavour compounds using a stable isotope
dilution assay. Lebensm. Wiss. Technol. 1990, 23, 513-522.
Kajiwara, T.; Harada, T.; Hatanaka, A. Isolation of Z-3-hexenal
in tea leaves, Thea sinensis, and synthesis thereof. Agric.
Biol. Chem. 1975, 39, 243-247.
Lin, J.; Arce Vera, F.; Welti, D. H.; Fay, L. B.; Blank, I.
Synthesis of deuterated volatile lipid degradation products
to be used as internal standards in isotope dilution assays.
1. Aldehydes. J. Agric. Food Chem. 1999, 47, 2813-2821.
Milo, C.; Blank, I. Quantification of impact odorants in food
by isotope dilution assay. Strength and limitations. In
Flavor AnalysissDevelopments in Isolation and Characterization; Mussinan, C. J., Morello, M. J., Eds.; ACS Symposium Series 705; American Chemical Society: Washington,
DC, 1998; pp 250-259.
J. Agric. Food Chem., Vol. 47, No. 7, 1999 2829
Mole, T.; Surtees, J. R. Reaction of monosubstituted acetylenes
with trialkylaluminums. Chem. Ind. 1963, 1727-1728.
Pritzkow, W.; Radeglia, R.; Schmidt-Renner, W. Studies on the
autoxidation of the isomeric n-octenes. J. Prakt. Chem. 1979,
321 (5), 813-826 (in German); Chem. Abstr. 1980, 92,
75520c.
Rakoff, H.; Rohwedder, W. K. Catalytic deuteration of alkynols
and their tetrahydropyranyl ethers. Lipids 1992, 27, 567569.
Rohwedder, W. K. Mass spectrometry of lipids labeled with
stable isotopes. Prog. Lipid Res. 1985, 24, 1-18.
Schieberle, P. New developments in methods for analysis of
volatile flavor compounds and their precursors. In Characterization of Food: Emerging Methods; Gaonkar, A. G., Ed;
Elsevier Science: Amsterdam, The Netherlands, 1995; pp
403-431.
Schieberle, P.; Grosch, W. Quantitative analysis of aroma
compounds in wheat and rye bread crusts using stable
isotope dilution assay. J. Agric. Food Chem. 1987, 35, 252257.
Snyder, E. I. Stereochemistry and mechanism of carboncarbon double-bond reduction in methyl cinnamate by
lithium aluminium hydride. J. Org. Chem. 1967, 32, 35313534.
Swoboda, P. A. T.; Peers, K. E. Metallic odour caused by vinyl
ketones formed in the oxidation of butterfat. The identification of octa-1,cis-5-dien-3-one. J. Sci. Food Agric. 1977, 28,
1019-1024.
Tressl, R.; Bahri, D.; Engel, K. H. Formation of eight-carbon
and ten-carbon components in mushrooms (Agaricus campestris). J. Agric. Food Chem. 1982, 30, 89-93.
Ullrich, F.; Grosch, W. Flavour deterioration of soya-bean oil:
Identification of intense odour compounds formed during
flavour reversion. Fat Sci. Technol. 1988a, 90, 332-336.
Ullrich, F.; Grosch, W. Identification of the most intense odor
compounds formed during autoxidation of methyl linolenate
at room temperature. J. Am. Oil Chem. Soc. 1988b, 65,
1313-1317.
van den Dool, H.; Kratz, P. A generalization of the retention
index system including linear temperature programmed
gas-liquid partition chromatography. J. Chromatogr. 1963,
24, 463-471.
Received for review February 19, 1999. Revised manuscript
received April 16, 1999. Accepted April 28, 1999.
JF9902090